Color energy as a seed descriptor for image segmentation with region growing algorithms on skin wound images

Misleading diagnosis of skin diseases may result in complications during the healing process. Skin images provide an important contribution to medical staff on storing and exchanging information to try preventing misdiagnosis. For such, image segmentation process may benefit from use of machine learning techniques, increasing simplicity of procedure, reducing computational costs and improving the diagnosis. This paper presents a comparison among different paradigms of machine learning to validate the segmentation of medical images of lower members ulcers, this segmentation allows wound pattern recognition to determinate injury region aiming at reducing the subjectivity of human evaluation.

show abstract

“…Figure 1a shows one of 33 images that were used to build the dataset files. Images acquisition and descriptions can were same as used in [6].…”

Section: Methodsmentioning

confidence: 99%

“…[6] uses colors to describe regions of the best representation of the lesion by analyzing colors present in a region, emphasizing the closest one to injury colors.…”

Section: Introductionmentioning

confidence: 99%

Pattern Recognition of Lower Member Skin Ulcers in Medical Images with Machine Learning Algorithms

Seixas

Barbon

Mantovani

2015

2015 IEEE 28th International Symposium on Computer-Based Medical Systems

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, in blue and green perspective (Figure 11c), cluster disjunctions are not consistent, thereby reducing overall precision. 661 In order to improve discrimination level between classes we can use other image descriptors, such as regions [25] [23] and differences between color channels [22]. K-means algorithm is one of the best known non-iterative clustering methods for unsupervised learning [6], but given the use of gold standard in our methodology, we could apply supervised learning modeling.…”

Section: K-means Clusteringmentioning

confidence: 99%

Robust computer vision system for marbling meat segmentation

Campos

Seixas

Barbon

et al. 2020

ELCVIA

View full text Add to dashboard Cite

In this study, we developed a robust automatic computer vision system for marbling meat segmentation. Our approach can segment intramuscular fat from meat samples using images acquired with different quality devices in an illumination varying environment, where there was external ambient light and artificial light; thus, professionals can apply this method without specialized knowledge in terms of image treatment or equipment, as well as without disruption to normal procedures, thereby obtaining a robust solution. The proposed approach for marbling segmentation is based on data clustering and dynamic thresholding. Experiments were performed using two datasets that comprised 82 images of 41 longissimus dorsi muscles acquired by different sampling devices. The experimental results showed that the computer vision system performed well with over 98% accuracy and a low number of false positives, regardless of the acquisition device employed.

show abstract

“…Seixas et al [9] implemented a segmentation approach for skin wound images. Their system is based on proposing an approach to find a seed for the region growing segmentation technique.…”

Section: Related Workmentioning

confidence: 99%

“…Second, most of the CW images are captured by using a regular camera with variable lighting conditions, which affects the wound images' quality [7]. Third, the assessment of the wound healing rate is considered a challenging task that depends on the segmentation of various tissue types [5], [9].…”

Section: Related Workmentioning

confidence: 99%

Chronic Wound Healing Assessment System Based on Different Features Modalities and Non-Negative Matrix Factorization (NMF) Feature Reduction

et al. 2019

View full text Add to dashboard Cite

Identification and treatment of chronic wounds (CWs) are considered economic and social challenges, especially with respect to bedridden and elderly persons. CWs do not follow a predictive course of healing within a particular period. Their treatment and management costs are very high. Also, CWs decrease quality of life for patients, which cause severe pain and discomfort. In this paper, we proposed a comprehensive wound healing assessment framework based on current appearance, texture, and prior visual appearance analysis to handle different types of CWs depending on extracting various tissue types. The framework provided an accurate evaluation tool for the CW healing process depending on extracting and fusing significant features from the CWs RGB images. Non-negative matrix factorization (NMF) was used to retrieve the most significant features to reduce computation time. The gradient boosted trees (GBT) classifier was used to classify different tissue types. Finally, the healing assessment of the CWs depended on calculating the improvement in the area of the necrotic eschar, slough, granulation, and healing epithelial tissues. The framework was trained and tested using 377 RGB images from Medetec wound database and national pressure ulcer advisory panel website. The proposed system achieved an average accuracy of 96% for tissue classification, which helps in obtaining an accurate CW healing assessment. This result can be considered as a promising result when compared to the other state-of-the-art techniques.INDEX TERMS Wounds and injuries, wound healing, texture analysis, color analysis, tissue segmentation.

show abstract

Color energy as a seed descriptor for image segmentation with region growing algorithms on skin wound images

Cited by 7 publications

References 26 publications

Pattern Recognition of Lower Member Skin Ulcers in Medical Images with Machine Learning Algorithms

Pattern Recognition of Lower Member Skin Ulcers in Medical Images with Machine Learning Algorithms

Robust computer vision system for marbling meat segmentation

Chronic Wound Healing Assessment System Based on Different Features Modalities and Non-Negative Matrix Factorization (NMF) Feature Reduction

Contact Info

Product

Resources

About